Microfiltration and ultrafiltration have been used for separation of compounds in biological broths or other liquids. The beverage industry has employed microfiltration to clarify beer and wine and in the dairy industry microfiltration and ultrafiltration can be used for processing of, for example, cheese whey or milk. Microfiltration has also recently been applied to the biotechnology industry, albeit somewhat more sparingly, for product separation and purification.
Microfiltration is in principle an attractive method of separating solutes from high solids suspensions, for example, fermentation suspensions, milk, or juice pulp. A variety of different microfiltration formats have been used in practice, including plate and frame, ceramic tubes, hollow fiber, and membrane systems. Plate and frame is used infrequently, but it is able to handle high solids concentrations. This format, however, is relatively expensive and requires a large equipment footprint when used for industrial scale operations. Ceramic tubes are widely used in the dairy and food industry because of the high throughputs, ease of operation, ease of sterilization/cleaning, and membrane longevity. However, ceramic tube systems are generally very expensive and require more power than other microfiltration systems in order to maintain the very high cross flows needed to minimize fouling. Hollow fibers are an alternative to ceramic tubes. They are not as operationally robust or as easy to run and operate as ceramic tubes, but are less costly and require a much smaller equipment footprint than ceramic tubes or plate and frame systems.
Spiral wound membranes have also been used for certain microfiltration operations. Spirally wound membrane constructions generally include an envelope of sheet membrane wound around a permeate tube that is perforated to allow collection of the permeate. Referring to FIG. 3, an exemplary spiral wound membrane module design includes a cylindrical outer housing shell, and a central collection tube sealed within the shell and having a plurality of holes or slots therein which serve as permeate collection means. A leaf comprising two membrane layers and a permeate channel layer sandwiched between the membranes is spirally wound around the tube with a feed channel spacer separating the layers of the wound leaf. The permeate channel layer typically is a porous material, which directs permeate from each membrane layer in a spiral path to the collection tube. In operation, a feed solution to be separated is introduced into one end of the cylinder and flows directly axially along the feed channel and feed spacer, and a retentate stream is removed from the other axial end of the shell. The edges of the membrane and permeate channel layer that are not adjacent the collection tube are sealed to retain and direct permeate flow within permeate channel layer between the membranes to the collection tube. Permeate which passes through the membrane sheets flows radially through the permeate collection means toward the central tube, and is removed from the central tube at a permeate outlet.
Applications of spirals at a commercial scale have been largely confined to treatments of highly dilute (low solids) process fluids. Spirally wound membrane modules are often employed alone or in combination for the separation of relatively low solids content materials by high pressure reverse osmosis, for example, for the production of pure water from brine; or low pressure ultrafiltration, for example, in the dairy field, for example, for the concentration of whey protein. In theory, a spiral wound membrane configuration offers a relatively large membrane surface area for separation processing relative to the footprint of the filtration module. The larger the membrane area in a filter system, the greater the permeation rate that is potentially available, everything else being equal. However, spiral wound membranes tend to foul at a high rate. Fouling leads to decline of flux, which determines system throughput, and decline in passage, which determines product yield. Unfortunately, the trans-membrane pressure (TMP) at the inlet of a spiral wound membrane is much higher than the TMP at the outlet. This occurs as membrane resistance creates a pressure gradient on the retentate side, whereas the permeate pressure is uniformly low across the membrane. Thus, optimal TMP condition can typically only be achieved within a relatively short zone along the membrane. Upstream of this optimal zone the membrane is overpressurized and tends to foul, while downstream of this zone the low TMP results in suboptimal flux. Spiral wound membranes are often run in series, which exacerbates the fouling problem.
Backpulsing is a generally known technique intended to restore flux and reduce fouling in filters. Backpulsing has been done in spiral membranes, for example, by forcing collected permeate backwards into the permeate channel to generate significant overpressure from the permeate side of the membrane. In the past, backpulsing strategies have not provided uniform local transmembrane pressures along the permeate side of the membrane. The pressure gradient within the permeate space has tended to be relatively higher at the permeate backflow inlet and relatively lower at distal locations in the permeate channel from the backflow source. Therefore, the level of localized defouling and flux restoration has varied considerably and unpredictably along the axial length of the membrane. In prior backpulsing approaches, either insufficiently low backflow pressure was developed within the permeate space resulting in suboptimal cleaning, or high backflow pressures developed within the permeate side sufficient to induce some level of defouling would lead to membrane damage by delamination. Backpulsing based on such permeate flow reversal techniques may generate a hydrodynamic shock wave or water hammer effect for inducing defouling, which is hard on the membrane. Also, the level of any flux restoration and defouling achieved tends to progressively decline after multiple filtration cycles using such backpulsing treatments. In some cases pressurized air has been used to enhance the backpulsing effect. However, some spiral membranes in particular may not be robust enough to tolerate pneumatic backpulsing. Some vendors, e.g., Trisep and Grahamtek, produce spiral membranes designed to handle backpulsing stresses.
Baruah, G., et al., J Membrane Sci, 274 (2006) 56-63, describe a microfiltration plant tested on transgenic goat milk featuring a ceramic microfiltration membrane configured with a back pulsing device, permeate re-circulation in co-flow to reportedly achieve uniform transmembrane pressure (UTMP), and a cooling/temperature control system. Backpulsing is done by trapping the permeate. This is done by closing the backpulse valve and a valve behind the pump outlet. By adjusting the bypass of the backpulsing device, a variable amount of liquid is then forced into the system to achieve the backpulse. However, modalities expected to cause non-uniform backpressure in the filtrate passage during backpulsing are undesirable as any defouling effects achieved on the membrane also will tend to be non-uniform. Also, ceramic filters generally are more costly than some other MF formats, for example, spiral membranes, and will offer less working surface area per length than a spiral format. Brandsma, R. L., et al., J Dairy Sci, (1999) 82:2063-2069, describe depletion of whey proteins and calcium by microfiltration of acidified skim milk prior to cheese making in a MF system reported to have UTMP capability. Alumina-based ceramic membranes are described as the filtering means, which were cleaned using a cycle of 1.5 weight percent NaOH and 1.5 weight percent nitric acid with use of the UTMP system as a backwashing mechanism. As such, the backwashing cycle as described by Brandsma et al. involves use of external chemicals to clean the ceramic membrane. The use of external harsh chemicals and significant production down times associated with their use to clean filters is non-ideal.
There is a need for filter strategies that can achieve high passage and yields in liquid/solid separations conducted on feed streams having low through high solids contents in a more continuous, less interrupted manner with reduced equipment and operating costs and effective defouling without cleaning chemical additions.
Crossflow filtration can also be used to separate like solutes or components based on differences in molecular weight. Sugar separation employing nanofiltration is one example. Separating milk proteins (primarily casein and whey) is another example that is actively being studied by the dairy industry. There has been some success with tubular ceramic membranes employing high crossflow velocities. Unfortunately, the hydrodynamics of spiral wound membranes have previously made this type of process very inefficient with polymeric spiral wound membranes, due to the development of a layer of polarized particles that eventually forms during operation. This fouling layer leads to reduced fluxes and rejection of solutes, specifically whey proteins. The fouling layer development is more extreme as the ratio between TMP and crossflow velocity increases. A system that can decouple crossflow from TMP would allow operation under conditions of minimal fouling